P-47 Maneuverability or Lack Thereof

[Legioneod]

2 hours ago, Denum said:

They peek in here every now and again.

 

Currently no one has really nailed down exactly what makes the P47 feel off, 

 

Could be the CL, could be elevator response etc. 

 

 

I'd say it's more than one issue but the problem is nailing them down and getting all the required data to prove it. I know that the P-47 in-game isn't producing proper power at altitudes and is using climb data instead of level flight data (from what I can see), which gives us around 200-300 less HP in certain altitudes/situations, and maybe even more so.

 

As far as Il2 vs DCS well imo DCS is the better modeled P-47 until someone proves otherwise.

[SCG_Tzigy]

1 hour ago, Holtzauge said:

There are numerous videos on Youtube showing cockpit views from WW2 fighters but one of my favourite ones showing the IRL “on rails” FM is this one with a Yak-3. Look at his stick input and how the plane responds rock solid. IMHO the closest thing we have in-game is a Fw-190A3 with 50% fuel. The in-game Spitfire MkXIV? I took that thing for a ride when it was released and now it’s just gathering dust in the hangar…

Great video! One really interesting thing jumped out at me, what is the very quick motion with the stick forward before each roll, is it a quick unload of the elevators? Not quite sure I saw that ever  before

[Holtzauge]

9 minutes ago, ZachariasX said:

This "mushiness around the center" is something I see on basically all aircraft and all simulators, this and that one as well. It just occurs more or less with each individual simulated aircaft. When they say that "this and that aircraft are very maneuvreable" then to me this implies that the plane acts swiftly upon stick input, meaning it is at least somewhat neutral in pitch axis, whereas on "less maneuvrable" aircraft, you really have to deflect that stick more and with more force, something that is not easily replicated in a sim with a table top joystick.

 

The Spitfire feels so maneuvrable because she feels like you could pull her through a loop with just one finger and on top of that her behaviour seems predictable and precise enough that makes you feel like pulling such a stunt. It has nothing to do with how fast the aircraft actually turns or rolls.

 

I think we are looking at a very basic simulation artifact here.

 

Yes, that we have to do some compromises since we don't "feel" the aircraft through the stick in a sim is also something that IMHO is also applicable to stick movement. It has become generally acceptable in the sim community not to have a 100% connection between the stick deflection on your desk and what goes on in the sim cockpit: A good example of this is the Me-109 roll and elevator response at high speeds: You can deflect your desktop stick all you like but nothing happens in "sim world" which is all good and fine because how else could you have it?

 

This brings me to the Spitfire: The purists will gasp and shudder if you were to suggest anything other than 100% realism on stick displacement: Move the stick 1/4" and the thing should buck. No matter that this creates big problems for controlling the thing in the sim. The solution to this could be to "filter" the desktop input in an intelligent way since this can't be addressed with stick settings since you get strange effects off-centre with non-linear effects or when staying linear with capping. It should be done intelligently in game just as it's done for force limitations. I know this will bug the purists but I would love to see them fly a F-16 with a realistic 1/16" stick deflection range. In a simulation you need to do compromises. However, whatever solution to this control problem is selected, the oscillating rubber band suspension FM should be taken out to pasture and the old "on rails" FM reinstated in all it's glory!

 

Grabs hat starts running......

[ZachariasX]

4 minutes ago, Holtzauge said:

Move the stick 1/4" and the thing should buck.

That seems actually true. The elevator has a rather long travel and the very subtle angular travel you give it is well in synch with the ailerons that are on a stick as short from the knee as a desktop stick is. It is the force required that throws the controls out of balance, not the differences in corresponding angular travel.

 

 

[Holtzauge]

10 minutes ago, SCG_Tzigy said:

Great video! One really interesting thing jumped out at me, what is the very quick motion with the stick forward before each roll, is it a quick unload of the elevators? Not quite sure I saw that ever  before

 

Did not notice that but the key takeaway I think is how crisply the aircraft responds to control input. There are other nice videos of WW2 warbirds where you can see a bit of the stick showing as well and where it's hard to detect any wobbliness: This one for a Me-109 for example.

[ACG_Cass]

My understanding is that IL2 employs a force based approach. So full stick is x amount of lbs of force. They also add in the delay of applying that force to the stick. It's a fairly decent way of doing things but when applied evenly across planes, can sometimes feel a bit off. 

 

Anyway. 

 

I've done some additional tests. This is incredibly difficult to test in the sim as the actual differences are quite minute across the planes. It's interesting to see how sensitive we are to these minute changes. 

 

The G6L and A5 were given the same relative fuel load as the D22 (~25%). Furthermore, in this application all 3 planes reached their full stick back position at almost exactly the same time and it lines up well with the maximum deflection point seen in the IRL test - I think we can take this to mean that stick forces aren't a factor at this speed. 

 

 

I also tested the P47 with full fuel and 2x1000lb bombs to see if that had any impact. The initially point of G onset and first few points were the same as the 11,900 lb. Obviously after that it starts to lag behind. 

 

What's difficult to understand is at what point does the DSV get created? (I'm certain at this point that it isn't something that is modeled in IL2 as none of the planes seem to show this behaviour) The reason I ask is to understand why the P47 is not matching the early pull we see above - is it because the DSV gets created very early and the additional lift means the G spikes quicker, or is there a lack of elevator authority in IL2 that means it doesn't have the responsiveness it has IRL?

 

I'd wager there is a big element of the latter. The P47 elevator is 22 sq. feet, which is absolutely massive. For reference:

 

P51 Take off Weight: 4578  Elevator: 13 sq feet - 351.6kg/sqft

P47 Take off Weight: 6213  Elevator: 22 sq feet - 282.4kg/sqft

 

Crude calculation but gives you an idea of the control surfaces the P47 had relative to it's weight. It really should be responsive and nimble up to it's Clmax. I'll need to research the 109 and 190 elevator surface areas before I can do further comparison but the idea that it could be more initially responsive than them doesn't seem too far fetched. 

 

@Holtzauge side point: what I didn't notice before in the IRL test is the elevator actually starts to reduce and the G still increases - I'm guessing that this is the DSV being created and adding the additional lift. 

 

 

Here is 190mph test:

 

[BCI-Nazgul]

On 12/19/2021 at 8:36 AM, Denum said:

The P47 D28 wants to stall at 300mph mid turn. Which seems odd. You can't even get that much stick deflection in before it tries to flip over on you.

That's one of my biggest complaints.  Your diving in at good speed in a turn and suddenly you're doing cartwheels even without a lot of stick deflection.  In DCS you would get a ton of warning before that happened.   In IL2 you get none.  I can understand that applying too much elevator too fast can cause an accelerated stall, but it seems way too abrupt with the 47.  Also, it lacks the ability to drag itself around slow turns using all that big engine power like it does in DCS.   In IL2 it just falls out of the sky.

On 12/19/2021 at 8:36 AM, Denum said:

The P47 D28 wants to stall at 300mph mid turn. Which seems odd. You can't even get that much stick deflection in before it tries to flip over on you.

That's one of my biggest complaints.  Your diving in at good speed in a turn and suddenly you're doing cartwheels even without a lot of stick deflection.  In DCS you would get a ton of warning before that happened.   In IL2 you get none.  I can understand that applying too much elevator too fast can cause an accelerated stall, but it seems way too abrupt with the 47.  Also, it lacks the ability to drag itself around slow turns using all that big engine power like it does in DCS.   In IL2 it just falls out of the sky.

On 12/19/2021 at 8:36 AM, Denum said:

The P47 D28 wants to stall at 300mph mid turn. Which seems odd. You can't even get that much stick deflection in before it tries to flip over on you.

That's one of my biggest complaints.  Your diving in at good speed in a turn and suddenly you're doing cartwheels even without a lot of stick deflection.  In DCS you would get a ton of warning before that happened.   In IL2 you get none.  I can understand that applying too much elevator too fast can cause an accelerated stall, but it seems way too abrupt with the 47.  Also, it lacks the ability to drag itself around slow turns using all that big engine power like it does in DCS.   In IL2 it just falls out of the sky.

[ACG_Cass]

144mph

 

 

This one looks particularly bad. 

[354thFG_Rails]

A3 is busted. I wouldn’t consider it to be close to how it should be. You might be right about the stick forces and it feeling like it’s on rails. But it can do some crazy things at the moment. That and the aircraft files seem off. The critical and danger AoA values are higher than all other 190’s. A3 currently is set to 17.5 and 20 degrees respectively. All other are set to 13.5 and 15.5 degrees. 
 

 

Edited December 20, 2021 by QB.Rails

[Legioneod]

6 minutes ago, QB.Rails said:

A3 is busted. I wouldn’t consider it to be close to how it should be. You might be right about the stick forces and it feeling like it’s on rails. But it can do some crazy things at the moment. 
 

 

I wonder if it has anything to do with the flaps. Flaps in Il2 seem to allow aircraft to do things that shouldn't happen. P-47 is a good example of this but I think all aircraft in Il2 are affected to some degree. Maybe the flap modeling on the whole needs a rework?

Edited December 20, 2021 by Legioneod

[354thFG_Rails]

5 minutes ago, Legioneod said:

I wonder if it has anything to do with the flaps. Flaps in Il2 seem to allow aircraft to do things that shouldn't happen. P-47 is a good example of this but I think all aircraft in Il2 are affected to some degree. Maybe the flap modeling on the whole needs a rework?

It very well could be flaps adding weird behavior but again I still think the A3 is far too maneuverable compared to the others. I understand different weights and all that. But the wing, from what I could find, did not change between the A and D models. 

[ACG_Cass]

Here are all the main aircraft tested under these conditions. All with ~25% fuel.

 

 

Here I've shifted the graph to line the P47 initial G point up with the IRL Test:

 

 

[ACG_Cass]

As far as I understand (and I may need to be corrected here), if a wing has a lower lift coefficient then pre-stall it would need more AoA in order to create the same load factor.

 

Now judging by the fact the 109, Typhoon, P51 and Spitfire track the line in the first image initially it seems as though the stick modelling is actually pretty spot on. I'd also guess that the DSV is created fairly quickly as the IRL P47 begins to leave them behind.

 

So in order for the P47 to keep up with them during that initial pull, it's elevators will need to outperform theirs in rotating the plane as it will need more AoA to create more lift. Its obviously on borrowed time as its Clmax is lower and it will be creating more drag, but judging by these tests the P47 elevators were pretty powerful.

 

A horrendous simplification that completely overlooks the intricacies of aerodynamics, but judging by weight vs elevator surface area, providing the P47s elevators were well designed this certainly seems feasible.

 

P47 - 245 kg  per sqft of elevator

109 - 307 kg  per sqft of elevator

Spit - 213 kg  per sqft of elevator

P51 - 309 kg  per sqft of elevator

 

(Still trying to find areas for the Anton and Typhoon)

 

Coupled with the fact that it has the largest propeller and most powerful engine this very could be the missing string from the IL2 47s bow that makes it such a dog.

 

I'll conduct some tests at the other speeds but judging just by the 144mph speed, something is clearly very off.

 

It could be that it's peak elevator authority is correct and that it takes too much deflection to get to those levels, but we'd need the devs to look into that as they are able to see things we can't. I'd also question that as in the 144mph test, it would have reached full deflection after 0.8s and you don't see the reaction you see IRL.

 

This also explains why the P47 feel so much nicer to fly with your trim mapped to your stick.

 

@Holtzauge

Probably gonna need a big fat sense check on the "logic" I've applied above ?

 

 

Edited December 20, 2021 by ACG_Cass

[ICDP]

3 hours ago, QB.Rails said:

A3 is busted. I wouldn’t consider it to be close to how it should be. You might be right about the stick forces and it feeling like it’s on rails. But it can do some crazy things at the moment. That and the aircraft files seem off. The critical and danger AoA values are higher than all other 190’s. A3 currently is set to 17.5 and 20 degrees respectively. All other are set to 13.5 and 15.5 degrees. 
 

 

 

What I find ironic is that you think the way the Yak 9 flew was in anyway much more "realistic".  I saw 2x aircraft being flown in utterly unrealistic ways and doing things neither should do in real life without serious consequences.

 

Dumping full flaps in seconds.

Full negative to positive G load in seconds

Being thrown around in ways that in real life would get you killed.

MP "aces" gaming the game to get a kill.

 

It is 100% why I stopped playing online in flight sims almost 20 years ago.  MP is just utter gash for this stuff, always has been and always will be.

Edited December 20, 2021 by ICDP

[Legioneod]

18 minutes ago, ICDP said:

 

What I find ironic is that you think the way the Yak 9 flew was in anyway more "realistic".  I saw 2x aircraft being flown in utterly unrealistic ways and doing thing neither should do in real life.

 

Dumping full flaps in seconds.

Full negative to positive G load in seconds

Being thrown around in ways that in real life would get you killed.

MP "aces" gaming the game to get a kill.

 

It is 100% why I stopped playing online in flight sims almost 20 years ago.  MP is just utter gash for this stuff, always has been and always will be.

Agreed. Only real way to combat this playstyle is to put more realistic limitations on the players that a real pilot would face. Pilot physiology was a step in the right direction imo. Next step imo would be limiting the player from doing certain things that wouldn't be possible irl (lowering flaps and pulling heavy g for example, or something to that effect) or even limiting the player to two hands, etc. 

I understand too much realism can be a bad thing but for a sim I don't think more realistic pilot limitations are too much to ask for, though I know some would find them unpopular.

It brings up an interesting question, if we want the aircraft to be realistic why not treat the pilot abilities/limitations in the same way?

Sorry for being OT.

Edited December 20, 2021 by Legioneod

[354thFG_Rails]

32 minutes ago, ICDP said:

 

What I find ironic is that you think the way the Yak 9 flew was in anyway much more "realistic".  I saw 2x aircraft being flown in utterly unrealistic ways and doing things neither should do in real life without serious consequences.

 

Dumping full flaps in seconds.

Full negative to positive G load in seconds

Being thrown around in ways that in real life would get you killed.

MP "aces" gaming the game to get a kill.

 

It is 100% why I stopped playing online in flight sims almost 20 years ago.  MP is just utter gash for this stuff, always has been and always will be.

Never brought up the yak because I was responding to the claim that the A3 at 50% is pretty close to real life “flying on rails”. I won’t post off topic things anymore sorry. 

[Denum]

1 hour ago, ICDP said:

 

What I find ironic is that you think the way the Yak 9 flew was in anyway much more "realistic".  I saw 2x aircraft being flown in utterly unrealistic ways and doing things neither should do in real life without serious consequences.

 

Dumping full flaps in seconds.

Full negative to positive G load in seconds

Being thrown around in ways that in real life would get you killed.

MP "aces" gaming the game to get a kill.

 

It is 100% why I stopped playing online in flight sims almost 20 years ago.  MP is just utter gash for this stuff, always has been and always will be.

?

 

The Yak wasn't doing anything in game that other aircraft aren't capable of. Everyone is confined to to same FM (mostly)

 

The majority, if not all of your complaints apply to every aircraft in the game. 

 

Realistic? Not entirely but we are getting pedantic here. It's still a game at the end of the day.

 

Alot of the MP aces are exceptional players regardless of what realism you add to it. Frankly they'd do even better if you made it more difficult. 

All that will achieve is punishing the average player. 

 

 

The A3 and Tempest are nearly UFOs in game. People are allowed to be frustrated by them. 

 

 

 

 

Edited December 21, 2021 by Denum

[Bert_Foster]

Flap Dumper cure .... Put a good DM in for Flap overspeed ... jammed or ripped off with attendant structural damage etc.

[Denum]

25 minutes ago, Bert_Foster said:

Flap Dumper cure .... Put a good DM in for Flap overspeed ... jammed or ripped off with attendant structural damage etc.

American flaps jam. 

 

Oddly enough the only ones in the game so far

[Holtzauge]

About the perceived trouble many are seeing with the P-47: Even if the onset of g’s in-game was to follow the NACA curve and not lag behind a bit as @ACG_Casshas shown, I don’t think this would make anyone happy because while the P-47 would then probably feel a bit more responsive, it would still not turn any tighter. So even if you would get to the limit a few 1/10 of a second faster and follow the NACA curve dead on, you will still not be able to reach the g-loads that people seem to think the P-47 should deliver. There are only two paths there as far as I can see:

 

One is an increase in the Clmax, This will bring two effects: One is a higher sustained turn rate at lower speeds than we see today. The second is a higher instantaneous turn rate. My impression is that this is what many people want to see. Another path to a higher sustained turn rate is if someone was to find evidence that power or propeller efficiency is missing in the current P-47 FM. In this case there will be no improvement in the instantaneous turn rate but the P-47 would have a slightly higher sustained turn rate at a slightly higher speed.

 

I could be wrong but given some of the comments that have been made so far, my impression is that when people say they want the P-47 to be more “maneuverable” they mean that it should turn tighter as in it should be possible to pull a bit harder turns, both sustained and instantaneous. If this is so then the only path there is to crank up the Clmax and I just can’t see there has been any evidence presented that supports this. Ample evidence for power on and abrupt stall Clmax has been presented but that does not address the core issue which is the P-47’s poor power off Clmax in comparison to the competition.

 

So summing up: IMHO a FM fix that fixes the lag in g-load onset would make the P-47 reach the limit a few 1/10 of a second faster would of course be nice but this would only address the first couple of seconds in a turn and you would still stall out as early as you do today and you would not be able to turn any tighter. If the latter is what’s really wanted, the only fix I can see is to dig up material supporting a higher power off Clmax which, given what we now know about the Republic S-3 airfoil, seems unlikely.

 

Finally just a comment on the Fw-190A3 FM: I have no idea if this crate is exceptional compared to the others in the extreme low speed part of the in-game FM. The emphasis here is on “game” since as @ICDPpointed out, this is just a game artifact. My comment to the “on-rails” FM was connected to the above stall speed part of the envelope and how the aircraft handles there. If the pilots in the Youtube videos I posted earlier on did something similar as @QB.Railsposted, it was probably redacted by Youtube for not fulfilling the “Do not try this at home!” rules.

 

 

 

 

Edit and PS @ACG_Cass: I don’t think the DSV is modeled in-game or even that it needs to be: This effect is IRL there for maybe a second or two. The important thing is the sustainable Clmax in power off conditions. Then regarding how you get there the Cl as a function of alfa (lift slope) is determined by the aspect ratio: The lower the aspect ratio, the flatter the lift slope is. The wing profile and setting to the datum line determines the Cl at zero alfa only. The lift slope from there on up to Clmax is set by the aspect ratio.

 

Edited December 21, 2021 by Holtzauge

[ACG_Cass]

3 hours ago, Holtzauge said:

About the perceived trouble many are seeing with the P-47: Even if the onset of g’s in-game was to follow the NACA curve and not lag behind a bit as @ACG_Casshas shown, I don’t think this would make anyone happy because while the P-47 would then probably feel a bit more responsive, it would still not turn any tighter. So even if you would get to the limit a few 1/10 of a second faster and follow the NACA curve dead on, you will still not be able to reach the g-loads that people seem to think the P-47 should deliver. There are only two paths there as far as I can see:

 

One is an increase in the Clmax, This will bring two effects: One is a higher sustained turn rate at lower speeds than we see today. The second is a higher instantaneous turn rate. My impression is that this is what many people want to see. Another path to a higher sustained turn rate is if someone was to find evidence that power or propeller efficiency is missing in the current P-47 FM. In this case there will be no improvement in the instantaneous turn rate but the P-47 would have a slightly higher sustained turn rate at a slightly higher speed.

 

I could be wrong but given some of the comments that have been made so far, my impression is that when people say they want the P-47 to be more “maneuverable” they mean that it should turn tighter as in it should be possible to pull a bit harder turns, both sustained and instantaneous. If this is so then the only path there is to crank up the Clmax and I just can’t see there has been any evidence presented that supports this. Ample evidence for power on and abrupt stall Clmax has been presented but that does not address the core issue which is the P-47’s poor power off Clmax in comparison to the competition.

 

So summing up: IMHO a FM fix that fixes the lag in g-load onset would make the P-47 reach the limit a few 1/10 of a second faster would of course be nice but this would only address the first couple of seconds in a turn and you would still stall out as early as you do today and you would not be able to turn any tighter. If the latter is what’s really wanted, the only fix I can see is to dig up material supporting a higher power off Clmax which, given what we now know about the Republic S-3 airfoil, seems unlikely.

 

Thanks @Holtzauge

 

Completely agree on the DSV aspect. I don't think it needs to be modelled in game. 

 

Aspect Ratios as I can find them are:

P47 - 5.61

P51 - 5.81

G14 - 6.06

 

So this would mean that the P47 has a flatter lift slope - would this mean that in order to match the rate of onset (only talking about initially here, not all the way to Clmax) of load factor, the P47 would need to induce AoA at a higher rate than both aircraft? Basically taking a single point from that graph, say at the 2g point, the P47 would need to be at X AoA, with the other aircraft lower. 

 

 

I do think you're underestimating the significance that the increased elevator authority would have. Those graphs are only for 1 second of time. The tests are very extreme pulls and at relatively low speed for instantaneous pulls (on the stats page the P47s sustained turn speed at 10,000 feet is 205mph and these tests were conducted at 6000 feet at 214 IAS) with relatively low power (27").

 

One of the main issues with the P47 in IL2 is its responsiveness. It's completely hopeless on the defensive as even if someone comes behind you with significantly more airspeed, they are able to turn inside you due to the lack of elevator authority. 

 

I still need to do some more tests but it appears that the P47 in DCS seems to have a similar stall speed to its IL2 counterpart in the same conditions and with the fact the IL2 P47 follows Manoeuvre Speed = Stall Speed x Square Root of Load Factor (G), I'd wager the DCS follows this as well. 

 

Looking at tests like this:

http://www.wwiiaircraftperformance.org/p-47/P-47_versus_FW-190.pdf

 

 

The P-47 is apparently at war load, which we will take as the D22s standard weight of 6213kg. Not even factoring in a use of fuel during the test, that puts the stall speed at 121mph for a Clmax of 1.23 (D22s according to the stats page). The minimum speed needed to get to 5g under these conditions is only 270mph and I would doubt they would be pushing much higher than that, certainly not in a sustained manner. 

 

This lack of elevator authority also explains why something like this isn't possible in game:

 

 

http://www.wwiiaircraftperformance.org/sl-wade.html

 

I know the Tempest isn't the best example to use in IL2, but almost any aircraft with a speed advantage of only 50mph can comfortably turn inside the P47. A lot of that is due to how slow the plane is to react in comparison. 

 

Another aspect to this would be the lack of induced drag from the elevators, which explains why the P47 keeps its speed better when trimmed fully rearward. I know the AoA will dwarf this from a drag perspective, but if you have 22sqft of airfoil (7.5% of the wing area) that needs to be moved to 2 degrees rather than 8, that's still going to have an impact. 

 

 

I'll give these a spin next as they are more gradual and won't include formation of a DSV. Obviously they aren't maximum performance turns so won't tell us about Clmax, but will at least give us an idea of speed loss.

 

No chance, so difficult to do this consistently to the point where it's comparable. 

 

Edited December 21, 2021 by ACG_Cass

[ACG_Cass]

Full trim, it does perform slightly better. I'd be very interested to see how the P47 would handle with elevators comparable to the IRL test. 

 

[Yak_Panther]

On 12/19/2021 at 3:18 AM, Holtzauge said:

Finally: Earlier on in this thread allegations about the reliability of results from airfoiltools.com were made with the implication that the data there is unreliable for the Republic S3 profile since it uses a too coarse panel mapping. However, the proper way (and intended) to use Xfoil is to use the coarser coordinate data file to create a finer mesh before running simulations. This has also been done when the L/D-polars for the S3 were generated at airfoiltools.  So doing your own simulations with the coarser coordinate data and based on that disqualifying the results is more a misunderstanding of how the data at airfoíltools was generated and how Xfoil should be used.

 The methods foiltools uses in plotting the polars is.

 

This method still results method discontinuities in the surface of the foil at the leading edge, because of the lack of data near the leading.  Using this method you still get a coarse pressure distribution at the leading edge. Even at low angles of attack. The result is that the pressure distribution over the leading edge of the airfoil still does not match wind tunnel data of a similar section.

 

 

Foil Tools S-3 airfoil with corrections applied.

 

 

 

The methodology utilized by foiltools results in a crooked camber line near the leading edge.  

The foiltools leading edge.

 

 

 

The Camber line as derived from the ords in NACA L4G12

 

 

Lets walk through the process FoilTools uses and see why we end up with crooked camber line at the leading edge. 

 

The foiltools methodology http://airfoiltools.com/polar/index.

 

First thing we do is import the ords in the correct format. 

 The set of ordinates foil tools lacks any data for first 1.25% of the foils upper surface. For ords xc .0125 to the origin, the software approximates the slope of the leading edge.  This introduces are first bit of error. The next refinement used by foil tools does not fix this lack of data. Nor does it match the leading edge shape of the real S-3-11.  

 

In step 2 foil tools used the MDES FILT tool to smooth out the profile.

 

We then re panel the foil to apply these changes to our rough foil. The default setting for Xfoil applies 160 panels to the new model.

 

This looks smoother, however if you look at the leading edge. It’s still coarse, the slope of the upper surface changes abruptly near where the original foil data ends at .0125xc.

 

This because methodology used by FoilTools is approximating the leading edge, linaerzering and smoothing. Part of why it’s not matching the pressure distribution of the real airfoil is because; Xfoil assumes the leading edge radius sits perpendicular to the chord line. Which is not the case for cambered airfoils. 

 

 

 

The result of all of this approximation is a camber line that wobbles near the leading. 

 

Because of the lack of data near the leading and methodology used by the website. The model isn’t representative of a S-3. Which is apparent when you compare the shape and magnitude of the pressure coefficients of the foil created by the foil tools method and the S-3 in similar conditions. 

 

MDES FLIT and pane cannot restore what information is missing from the foil data. Using the foiltools method is still produces inaccurate data. The problem with the suit of tools in Xfoils is that if your not cognizant of the underlying  foil geometry it's vary easy to distort the airfoil beyond the geometry which actually defines the airfoil.  The thickness ratio and the leading radius are just one of the many features that define an airfoil.

 

The Republic S-3 has an different thickness distribution than the NACA 23000 series. The thickness distribution is given as.

 

Using the leading edge radius as means to evaluate Cl max of two different airfoils with different thickness distributions is misleading. The leading edge radius is optimized for the thickness distribution of the airfoil and the slope of the camber at the leading edge. If the LE radius is to big thickness distribution or the slope of the lead edge radius it can hurt performance. So you just can't say airfoil A will have better lift performance the B foil because A has larger leading edge radius.     

From NACA 824

[ACG_Cass]

Turns out my 144mph test was wrong. I overlayed it over the 190mph chart. 

 

Updated here with a Spit IX to compare to. 

 

 

Interesting part is if you line up the chart right at the start of the force being applied, the Spitfire lines up pretty spot on. I'm unsure what the delay is on force applied to the stick. I'm sure IL2 has at least some so it isn't out of the question that the ramp up of force is close to this IRL stick pull example. 

 

 

Does this mean the elevators for the P47, would allow it to almost match a Spitfire IXs pull up to the point of stall? It looks like it stalls at about 1.7g. 

 

My hunch is that the max elevator force for the P47 is correct, but forces below that seem to drop off too quickly. 

[ACG_Cass]

DCS Isn't going to be possible to test under the same scenario I don't think. As IL2 doesn't have a G-Meter in the replay, I'm not able to play something at full speed and then capture that. 

 

In DCS, Stick position is much more 1-1 and is not affected by slowing down time. At 1/32 the stick just flies straight back. 

 

What is interesting is comparing the P51 and P47 in DCS under the test scenarios. 

 

 

This looks weird as the P47 is able to pull higher G. But going off the theory that the P47s elevators are much stronger - is it possible that the P47 is creating a stronger DSV that is allowing for more lift?

 

Fortunately DCS has an AoA meter:

 

As you can see the P47s elevators are massively out performing the P51s in the rotation of the plane. 

 

Sim to Sim comparison wise it shows that IL2s P47 elevators are way under what they should be. 

 

 

I'll continue to do some tests in DCS, but I already have a hunch that the Clmax both sims have reached won't be miles off. The biggest disparity is the lackluster elevators. If you're able to perform a maneuver with 2 degrees of elevator rather than 8, I think that would still affect the drag profile of the aircraft fairly significantly.  

 

 

[Holtzauge]

1 hour ago, Yak_Panther said:

 The methods foiltools uses in plotting the polars is.

 

This method still results method discontinuities in the surface of the foil at the leading edge, because of the lack of data near the leading.  Using this method you still get a coarse pressure distribution at the leading edge. Even at low angles of attack. The result is that the pressure distribution over the leading edge of the airfoil still does not match wind tunnel data of a similar section.

 

 

Foil Tools S-3 airfoil with corrections applied.

 

 

 

The methodology utilized by foiltools results in a crooked camber line near the leading edge.  

The foiltools leading edge.

 

 

 

The Camber line as derived from the ords in NACA L4G12

 

 

Lets walk through the process FoilTools uses and see why we end up with crooked camber line at the leading edge. 

 

The foiltools methodology http://airfoiltools.com/polar/index.

 

First thing we do is import the ords in the correct format. 

 The set of ordinates foil tools lacks any data for first 1.25% of the foils upper surface. For ords xc .0125 to the origin, the software approximates the slope of the leading edge.  This introduces are first bit of error. The next refinement used by foil tools does not fix this lack of data. Nor does it match the leading edge shape of the real S-3-11.  

 

In step 2 foil tools used the MDES FILT tool to smooth out the profile.

 

We then re panel the foil to apply these changes to our rough foil. The default setting for Xfoil applies 160 panels to the new model.

 

This looks smoother, however if you look at the leading edge. It’s still coarse, the slope of the upper surface changes abruptly near where the original foil data ends at .0125xc.

 

This because methodology used by FoilTools is approximating the leading edge, linaerzering and smoothing. Part of why it’s not matching the pressure distribution of the real airfoil is because; Xfoil assumes the leading edge radius sits perpendicular to the chord line. Which is not the case for cambered airfoils. 

 

 

 

The result of all of this approximation is a camber line that wobbles near the leading. 

 

Because of the lack of data near the leading and methodology used by the website. The model isn’t representative of a S-3. Which is apparent when you compare the shape and magnitude of the pressure coefficients of the foil created by the foil tools method and the S-3 in similar conditions. 

 

MDES FLIT and pane cannot restore what information is missing from the foil data. Using the foiltools method is still produces inaccurate data. The problem with the suit of tools in Xfoils is that if your not cognizant of the underlying  foil geometry it's vary easy to distort the airfoil beyond the geometry which actually defines the airfoil.  The thickness ratio and the leading radius are just one of the many features that define an airfoil.

 

The Republic S-3 has an different thickness distribution than the NACA 23000 series. The thickness distribution is given as.

 

Using the leading edge radius as means to evaluate Cl max of two different airfoils with different thickness distributions is misleading. The leading edge radius is optimized for the thickness distribution of the airfoil and the slope of the camber at the leading edge. If the LE radius is to big thickness distribution or the slope of the lead edge radius it can hurt performance. So you just can't say airfoil A will have better lift performance the B foil because A has larger leading edge radius.     

From NACA 824

 

@Yak_Panther: Maybe the best course of action for you would have been to just let this go? Getting on a high horse now to explain how Xfoil "works" and its “limitations” would have carried more water if you had not made this post (complete with pictures) before.

 

Especially this is telling: “The lift and drag curves on Foiltools.com  for the Republic S-3 are not reliable. The model lacks enough detail for comparative analysis. There are far too few panels used in their modeling. There are only 37 panels on their S3 airfoil. The result is that the leading edge geometry is far too coarse.”

 

Then: “The result is premature shearing of flow and a nonlinear distribution of pressure along the leading edge.”

 

Of course there is! The panels are simply too few in your model! You even posted a zoomed up picture of the nose and pressure distribution for the 37 panel model which you made which of course produces nonsense results but that’s the way simulators work: Garbage in- garbage out. As I already explained in a previous post: The airfoiltools Republic S-3 polars were NOT done using the database 37 panels like you claimed in the linked post above.

 

Now you are back posting pictures of nose panelings for the S-3 profile which look like they have 160 or above panels and consequently starting to learn a thing or two about Xfoil. However, doing a longwinded post now going into second order effects of the camber line when having done a huge first order mistake on paneling does not hold water.

 

If you had just winded your neck in a bit I would have let this pass but you didn’t.  The only reason I’m posting this now is because you persist in posting stuff here posing as an expert when you have no idea about the subject which your take on Xfoil and the mix of NACA charts for power on, power off and abrupt stall conditions shows. But the presentation is certainly impressive. This was why I came to think of the Retro Encabulator actually.

 

And now you continue with “Lets walk through the process FoilTools uses” & "The problem with the suit of tools in Xfoils is" etc. Talk about the blind leading the blind. But the amusement factor is high: Here we have an internet “expert” explaining the shortcomings of a simulation tool developed by MIT professor Mark Drela, a widely respected and absolute icon in the aerodynamic community.

[Yak_Panther]

2 hours ago, Holtzauge said:

 

If you had just winded your neck in a bit I would have let this pass but you didn’t.  The only reason I’m posting this now is because you persist in posting stuff here posing as an expert when you have no idea about the subject which your take on Xfoil and the mix of NACA charts for power on, power off and abrupt stall conditions shows. But the presentation is certainly impressive. This was why I came to think of the Retro Encabulator actually.

 

And now you continue with “Lets walk through the process FoilTools uses” & "The problem with the suit of tools in Xfoils is" etc. Talk about the blind leading the blind. But the amusement factor is high: Here we have an internet “expert” explaining the shortcomings of a simulation tool developed by MIT professor Mark Drela, a widely respected and absolute icon in the aerodynamic community.

The simulation aspect of the tool isn't at issue. It's how the user, you and foiltools, are using the geometric tools without any regard to the underlying geometry of the foil. Then purport the result as infallible. The type of spline doesn't work for a few reasons at it's obvious.  First is base panel coarseness which I've already covered.   Stuffing 300 panels into the model doesn't fix the core geometric problems with the foil.

  

Nor does it result in a model which matches the pressure distribution of the real thing.

 

The Second reason why the ords on foiltools won't match a windtunnel test of the airfoil is; The foiltool ordinates are from a wing with pretty large surface irregularities.

 

If you you knew what you were doing with the tool. those irregularities would have been apparent from the get.  You then could apply the appropriate corrections and mitigate some of the issues.  Or plot the ordinates and attempt to resolve surface imperfections through regression. Maybe even read NACA L4G12 and extract the ords from the S-3 model in that report and compare the foil geometry with ords in NACA L6B21. Then create a new set of ords to test with.

 

 

 

but yeah I'm the one who doesn't know what I'm doing...

Edited December 21, 2021 by Yak_Panther

[Holtzauge]

10 minutes ago, Yak_Panther said:

but yeah I'm the one who doesn't know what I'm doing...

 

Yes exactly: You are the one who does not know what you are doing. From a post you make October 2:

 

“I dug through the NASA technical reports about the P-47. Based upon my research the lift coefficient for the P-47 game looks low, the elevator's effectiveness, (ability to generate pitch rate),  and the increased loss of elevator effectiveness due to air speed / mach effect is too high. In short, The P-47 it should be more responsive to initial pitch inputs. I will explain each issue in detail below. 

 

 

            CL Max

 

 

            The Cl max, in the game, is probably low. It’s currently 1.2 according to the game info. Based on NACA reports the power on Cl max of the P-47 is around 1.4 - 1.5. “

 

Here you are in the same sentence mixing the power off Clmax we have in game which is 1.2 with the power on Clmax which clear as day shows you have no clue.

 

In the same post you also include NACA charts with abrupt pull-out stall Clmax in the context of the in-game Clmax being 1.2 in power off 1g stall conditions which is yet again comparing apples with pears. I posted data earlier on showing why you cannot connect the two. So yes,  you are clueless.

 

Packaging all that into long posts with Rockwell Retro Encabulator talk spiced with a generous helping of NACA charts does not change that fact. It only makes it all the more annoying.

 

I'm not wasting any more time on this. Welcome to my ignore list.

 

[Bremspropeller]

6 hours ago, ACG_Cass said:

The P-47 is apparently at war load, which we will take as the D22s standard weight of 6213kg. Not even factoring in a use of fuel during the test, that puts the stall speed at 121mph for a Clmax of 1.23 (D22s according to the stats page). The minimum speed needed to get to 5g under these conditions is only 270mph and I would doubt they would be pushing much higher than that, certainly not in a sustained manner. 

 

What does the snippet really say?

The P-47 turned better, but the 190 was heavy in pitch-control, but tended to black-out the pilot? How is that even possible?

The next paragraph has the 190 (that allegedly climbed worse) turn tighter at slow speeds and finds itself accelerating quicker and "easy to convert to a better position".

 

The report is way too unspecific about what it's actually describing for using it as more than anecdotal data.

[ZachariasX]

16 minutes ago, Bremspropeller said:

The P-47 turned better, but the 190 was heavy in pitch-control, but tended to black-out the pilot? How is that even possible?

When the 190 is actually going faster than the P47 and gives you more pitch authority at higher speeds without the pilot realizing. Then you have your heavy stick force on the 190 AND you black out as the report states. Also you have the P47 doing a smaller circle. The rest of the description is rather in line with the P47 being dogfood for the 190, unless the engagement is very high up or the P47 has initial advantage.

[Legioneod]

2 minutes ago, ZachariasX said:

When the 190 is actually going faster than the P47 and gives you more pitch authority at higher speeds without the pilot realizing. Then you have your heavy stick force on the 190 AND you black out as the report states. Also you have the P47 doing a smaller circle. The rest of the description is rather in line with the P47 being dogfood for the 190, unless the engagement is very high up or the P47 has initial advantage.

10,000 ft isn't very high imo. Also I wouldn't say the P-47s were dogfood for 190s down low as many pilot accounts state otherwise. Many said 190s were never a real problem at any altitude.

Either way the test and pilot accounts aren't going to give us the info we really need on determining the P-47s proper capabilities. They certainly interesting and give an idea of what was possible but still not useful in this case.

 

There is guncam footage out there with a 190 engaging a P-47 in a turn fight down low. Neither one was really able to outturn the other, and they were pretty much on par with each other for the most part.

 

Hopefully someone can find the data to show what it should be (if they haven't already)

I'm curious to know why DCS feels so different vs IL2 when it comes to the P-47. They feel like completely different aircraft to me personally.

I don't really know much, but if DCS and Il2 have similar stall speeds, etc then why such a difference in performance/feel overall?

[Holtzauge]

8 hours ago, ACG_Cass said:

 

Thanks @Holtzauge

 

Completely agree on the DSV aspect. I don't think it needs to be modelled in game. 

 

Aspect Ratios as I can find them are:

P47 - 5.61

P51 - 5.81

G14 - 6.06

 

So this would mean that the P47 has a flatter lift slope - would this mean that in order to match the rate of onset (only talking about initially here, not all the way to Clmax) of load factor, the P47 would need to induce AoA at a higher rate than both aircraft? Basically taking a single point from that graph, say at the 2g point, the P47 would need to be at X AoA, with the other aircraft lower. 

 

 

I do think you're underestimating the significance that the increased elevator authority would have. Those graphs are only for 1 second of time. The tests are very extreme pulls and at relatively low speed for instantaneous pulls (on the stats page the P47s sustained turn speed at 10,000 feet is 205mph and these tests were conducted at 6000 feet at 214 IAS) with relatively low power (27").

 

One of the main issues with the P47 in IL2 is its responsiveness. It's completely hopeless on the defensive as even if someone comes behind you with significantly more airspeed, they are able to turn inside you due to the lack of elevator authority.

 

Yes, the lift slope is very much connected to the aspect ratio but the difference between the planes you listed is not huge and if you google a bit on aspect ratio you will find good explanations of the effect. It does have an effect on the g-rate onset as you say though: In fact when the JAS 39 Gripen was being promoted by SAAB to the SwAF, the SAAB marketing people turned a lower aspect ratio compared to the F-16 (meaning more induced drag in turn) into an "advantage" by highlighting that the Gripen's susceptibility to turbulence and gusts when flying low level attack missions would make it a smoother ride for the pilot!

 

About underestimating the importance of elevator authority then maybe: It depends on what you mean? If it's responsiveness then I'm on the same page but my point is still this: Even if you get there (to a high g-load) a little faster how does it help you? If this takes 1.3 or 1.5 seconds because once you get there, you will still be limited by a power off stall Clmax of 1.2 for the duration of the turn which is probably 10 or multiples of 10 seconds?

 

Concerning the bleed of airspeed in a tight turn, this can for a first order approximation be assumed to be independent of profile drag because the overwhelmingly large part of the drag in a turn is the induced, not the increased profile drag, so your speed bleed will depend on the Cl you can attain and the aspect ratio.

 

But for sure: If you don't have enough elevator authority to either reach Clmax, or reach it in a sufficiently low time then you of course have a problem. However, in a turn held for 10 to 20 s, is a few 1/10 of a second to reach Clmax really an issue?

 

1 hour ago, Bremspropeller said:

 

What does the snippet really say?

The P-47 turned better, but the 190 was heavy in pitch-control, but tended to black-out the pilot? How is that even possible?

The next paragraph has the 190 (that allegedly climbed worse) turn tighter at slow speeds and finds itself accelerating quicker and "easy to convert to a better position".

 

The report is way too unspecific about what it's actually describing for using it as more than anecdotal data.

 

55 minutes ago, ZachariasX said:

When the 190 is actually going faster than the P47 and gives you more pitch authority at higher speeds without the pilot realizing. Then you have your heavy stick force on the 190 AND you black out as the report states. Also you have the P47 doing a smaller circle. The rest of the description is rather in line with the P47 being dogfood for the 190, unless the engagement is very high up or the P47 has initial advantage.

 

Yes, this is the problem with anecdotal stuff I think: What does it mean really and how to interpret it? My best example is Erwin Leykauf's claim that he was always able to out-turn Spitfires in his Me-109. And he probably did this as well! However, is that representative of the aircraft or is it an indication of his skill versus his opponents lack of skill? I think we all know the answer too that one but this highlights the problem with anecdotal evidence I think: Some of it is not really representative but OTOH, if you add up a lot of pilots opinions they will probably mostly agree that the Spitfire DID out-turn the Me-109 so anecdotal evidence is of course helpful once you have gotten rid of the outliers. Problem is if you are sitting there with just one or two stories? Are they representative or are they outliers like the Leykauf story?

 

Again the amazing merge of posts! Brems liked my first post but now I've gone and added some more stuff which is on a completely separate topic so this means he has "automatically liked what I just added which I intended to post in a separate post.

 

Any moderator who would care to comment on this? What's the plan? When is this going to be fixed!!!!!!

Edited December 21, 2021 by Holtzauge

[Legioneod]

20 minutes ago, Holtzauge said:

 

Yes, the lift slope is very much connected to the aspect ratio but the difference between the planes you listed is not huge and if you google a bit on aspect ratio you will find good explanations of the effect. It does have an effect on the g-rate onset as you say though: In fact when the JAS 39 Gripen was being promoted by SAAB to the SwAF, the SAAB marketing people turned a lower aspect ratio compared to the F-16 (meaning more induced drag in turn) into an "advantage" by highlighting that the Gripen's susceptibility to turbulence and gusts when flying low level attack missions would make it a smoother ride for the pilot!

 

About underestimating the importance of elevator authority then maybe: It depends on what you mean? If it's responsiveness then I'm on the same page but my point is still this: Even if you get there (to a high g-load) a little faster how does it help you? If this takes 1.3 or 1.5 seconds because once you get there, you will still be limited by a power off stall Clmax of 1.2 for the duration of the turn which is probably 10 or multiples of 10 seconds?

 

Concerning the bleed of airspeed in a tight turn, this can for a first order approximation be assumed to be independent of profile drag because the overwhelmingly large part of the drag in a turn is the induced, not the increased profile drag, so your speed bleed will depend on the Cl you can attain and the aspect ratio.

 

But for sure: If you don't have enough elevator authority to either reach Clmax, or reach it in a sufficiently low time then you of course have a problem. However, in a turn held for 10 to 20 s, is a few 1/10 of a second to reach Clmax really an issue?

 

Would a dive make any difference? Are we limited by the same things? I don't really know enough but for me the elevator authority is more important in a dive than in a turn (at least for the P-47). The ability to pull lead quickly in a dive can make the difference between getting a shot off vs missing the target (I guess the same thing could be said for turns as well)

Edited December 21, 2021 by Legioneod

[354thFG_Rails]

2 minutes ago, Legioneod said:

Would a dive make any difference? Are we limited by the same things? I don't really know enough but for me the elevator authority is more important in a dive than in a turn (at least for the P-47). The ability to pull lead quickly in a dive can make the difference between getting a shot off vs missing the target (I guess the same thing could be said for turns as well)

I think the point being made from what I can gather on the elevator pitch response is that it’s delayed enough causing enough drag to where it can’t pull as hard at a certain speed because it’s slowed down to much. From what Cass has shown it looks like the 47 should be pitching quicker with less input from the controls. 

[Holtzauge]

23 minutes ago, Legioneod said:

Would a dive make any difference? Are we limited by the same things? I don't really know enough but for me the elevator authority is more important in a dive than in a turn (at least for the P-47). The ability to pull lead quickly in a dive can make the difference between getting a shot off vs missing the target (I guess the same thing could be said for turns as well)

 

If you mean pulling a tighter turn then you would still be limited by the Clmax when diving slightly to gain energy but again, if you feel you can't stall the plane properly then the elevator authority is of course a problem but as long as you have to hold back a bit to avoid stalling, then the Clmax is still the culprit.

 

19 minutes ago, QB.Rails said:

I think the point being made from what I can gather on the elevator pitch response is that it’s delayed enough causing enough drag to where it can’t pull as hard at a certain speed because it’s slowed down to much. From what Cass has shown it looks like the 47 should be pitching quicker with less input from the controls. 

 

If you truly are being slowed down in-game by the tailplane related drag that would be an issue for sure but I would be surprised if you did see any speed change due to the drag caused by the elevator being deflected: This controls related drag will "drown" in the overwhelmingly biggest part which will still be the induced drag from the wings caused by taking out more lift from them.

Edited December 21, 2021 by Holtzauge

[Legioneod]

18 minutes ago, Holtzauge said:

 

If you mean pulling a tighter turn then you would still be limited by the Clmax when diving slightly to gain energy but again, if you feel you can't stall the plane properly then the elevator authority is of course a problem but as long as you have to hold back a bit to avoid stalling, then the Clmax is still the culprit.

 

If you truly are being slowed down in-game that would be an issue for sure but I would be surprised if you did see any speed change due to the drag caused by the tailplane and elevator being deflected: This drag will "drown" in the overwhelmingly biggest part which will still be the induced drag from the wings caused by taking out more lift from them.

I don't mean turning at all, just a straight diving attack. The ability to pull lead quickly is important in this situation imo. If I can't pull enough quickly enough or as much as I should be able, then it affect my ability to attack. Would the things Cass is talking about have an effect in this situation?

[ACG_Cass]

1 hour ago, Bremspropeller said:

What does the snippet really say?

The P-47 turned better, but the 190 was heavy in pitch-control, but tended to black-out the pilot? How is that even possible?

The next paragraph has the 190 (that allegedly climbed worse) turn tighter at slow speeds and finds itself accelerating quicker and "easy to convert to a better position".

 

The report is way too unspecific about what it's actually describing for using it as more than anecdotal data.

You're right it is too unspecific to be used just by itself as evidence, but combined with everything else it had the potential to be relevant. The report clearly condenses a lot of different scenarios into a paragraph so it's difficult to dissect. What I was trying to get across was that above 250mph, elevator authority becomes a factor and what we can see from IRL tests is the P47 had a lot of it. The mention of heavy controls for the 190 could have been the inability of the pilot to pull with the P47 - but honestly there is no way of deducing anything concrete from that test by itself. 

 

 

[Legioneod]

3 minutes ago, Denum said:

In my experience so far the 47 does lose an overwhelming amount of speed in turns. 

 

I fully expect to lose a turnpike on the back to a 109, but when they're able to outturn me and pull a thousand yards of distance on me because they maintain their speed so well through a turn little tough to understand.

 

 

 

 

This is another thing I'm curious about. Was the drag really so bad on the P-47 that it loses speed at the drop of a hat? I've read or watched a video somewhere stating the drag on the 47 was on par (or not much worse) in comparison to other aircraft so I'm not sure why it losses it's speed so quickly. Also its inability to maintain energy is very frustrating.

Edited December 21, 2021 by Legioneod

[Holtzauge]

1 minute ago, Denum said:

In my experience so far the 47 does lose an overwhelming amount of speed in turns. 

 

I fully expect to lose a turnpike on the back to a 109, but when they're able to outturn me and pull a thousand yards of distance on me because they maintain their speed so well through a turn little tough to understand.

 

 

 

 

What you describe (the Me-109 maintaining speed better) is due to the P-47 having a higher span loading than the Me-109 so the P-47 should slow down faster. However, note I'm not saying that it's correct the way it is now: Could be there is something off and it IS slowing to fast in-game but this would be very difficult to prove even if it was so I think. That it out-tuns you is again due to wing loading and Clmax.

[ACG_Cass]

55 minutes ago, Holtzauge said:

Concerning the bleed of airspeed in a tight turn, this can for a first order approximation be assumed to be independent of profile drag because the overwhelmingly large part of the drag in a turn is the induced, not the increased profile drag, so your speed bleed will depend on the Cl you can attain and the aspect ratio.

 

But for sure: If you don't have enough elevator authority to either reach Clmax, or reach it in a sufficiently low time then you of course have a problem. However, in a turn held for 10 to 20 s, is a few 1/10 of a second to reach Clmax really an issue?

My point was these are very extreme tests. You would never find yourself pulling 100% stick at 214 IAS as you'd very quickly stall, as we see. I can test at higher speed but my guess would be that the rate of onset differences will be more pronounced there and a difference of a few tenths or half a second head start to a turning rate can be the difference between life and death. Especially when approaching 7g cornering speeds.  

 

My question would be, outside of maximum performance turns would profile drag be a factor? You aren't always doing tight turns and the P47 bleeds speed in even the lightest of manoeuvres.